Mixed Method Data Collection Strategies (review)

There have been many recent considerations of survey methods, ethnographic methods, focus group methods, archival methods, quantitative methods and qualitative methods. The authors of this book assume that readers will be familiar with the logic and practice of those individual methods. The aim here is to make the case that the quality of social research can be improved through the use of mixed method data collection strategies. This pushes us beyond mere tolerance of methodological diversity to the idea that serious social research often must make use of mixed methods. A strength-of-methodological-diversity argument is not new, as it echoes themes in Stanley Lieberson's 1991 presidential address to the American Sociological Association and exhortations by the renowned demographer John Caldwell, two of many voices in this broader movement. In this book, William G. Axinn and Lisa D. Pearce succeed in further developing and illustrating the strength-of-methodological-diversity argument.

The book can be split into two parts, one general, one specific. In chapters 1 ("Motivations for Mixed Method Social Research"), 2 ("Fitting Data Collection Methods to Research Aims"), and 8 ("Conclusion"), the authors review and refine the general case for using more than one method of data collection. Some parts of the discussion may seem familiar to readers. For example, two pillars of the argument are the notions of counterbalance and comprehensiveness. All methods have strengths and weaknesses, but one method's strength is often another method's weakness. The combination of highly structured survey methods with less structured interviewing or observational methods provides counterbalance as it facilitates both the testing and discovery of hypotheses and causal mechanisms. Likewise, the authors argue that methodological diversity yields a more comprehensive empirical record. That record is more complete, in that it captures more aspects of any given social process or pattern, as well as more redundant, in that it includes overlapping measures provided by different methods. Readers familiar with triangulation and more general calls for methodological diversity will likely view this as an accessible and useful review. But other aspects of the argument are fresher and quite intriguing. The authors argue that new insights into cause and consequence in the social sciences are based on a constant interplay between theory and evidence. Better evidence, obtained through multiple methods, advances causal reasoning. As investigators, our distance from evidence is also important, as introspection is enhanced and guided by direct involvement in data collection and debilitated and distorted by distance. These and other arguments, developed in some detail, add new ideas to an otherwise competent review.

In the balance of the book (chapters 3–7), the authors provide concrete examples of mixed method data collection strategies from their demographic research, mostly from a project conducted in Nepal. These chapters describe how and why the authors employed mixed methods. They also describe in detail a few valuable methodological innovations and illustrate how mixed methods of data collection improved the research. Innovations include the "micro-demographic community study approach" (Chapter 3), "systematic anomalous case analysis" (Chapter 4), and two types of "history calendars" that facilitate and improve the collection of longitudinal data on neighborhoods (Chapter 5) and individuals (Chapter 6). Sprinkled throughout these chapters are compelling examples of how the simultaneous use of multiple methods generated fresh insights into social processes. For example, a combined ethnographic and survey approach allowed them to learn of the importance of a small development project, the Small Farmers Development Program, and include it in a survey about contraceptive use. Logistic regression analyses show that the omission of an indicator of SFDP participation leads to a notable downward bias in the estimate of the effect of wage work on contraceptive use. The moral of this and other examples is that mixed methods of data collection minimize bias and advance causal reasoning through the discovery of new causal mechanisms.

The book's strengths also reveal some weaknesses. No doubt, the best way to illustrate the strength-of-methodological-diversity argument is...